Skip to main content English


Two SAS macro programs are presented that compute the relative importance of covariates in the proportional hazards regression model and in the logistic regression model (Heinze & Schemper, 2003). The relative importance of covariates is quantified by the proportion of explained variation in the outcome (PEV) attributable to those covariates.
For proportional hazards regression, the program RELIMPCR uses the recently proposed measure V (Schemper & Henderson, 2000) to calculate the proportion of explained variation. RELIMPCR now replaces the earlier macro RELIMP (Schemper, 1993) which compared Maddala's likelihood-based R-squared measure (Maddala, 1983).
For the logistic model, the R-squared measure based on Pearson residuals is used by the program RELIMPLR (Mittlböck & Schemper, 1996).
Both programs are able to compute marginal and partial PEV, to compare PEV between groups of variables, and even to compare PEV between different models. They use a bootstrap resampling scheme to assess the covariance of PEV of different factors or models. Confidence limits for P-values are provided. The programs further allow to base the computation of PEV on models with shrinked (Copas, 1997) or bias-corrected (Heinze & Schemper, 2001, 2002) parameter estimates. Methods used by the programs and the usage of the macros are documented in a Technical Report


Heinze, G., Schemper, M. (2012): "RELIMPCR and RELIMPLR. SAS-macros for the analysis of relative importance of prognostic factors in Cox and logistic regression" Technical Report 4/2012, Section for Clinical Biometrics, CeMSIIS, Medical University of Vienna, Vienna
Dunkler D., Michiels S, Schemper M. (2007): "Gene expression profiling: Does it add predictive accuracy to clinical characteristics in cancer prognosis?" European Journal of Cancer 2007; 43(4): 745-751.
Heinze, G., Schemper, M. (2003): "Comparing the importance of prognostic factors in Cox and logistic regression using SAS", Computer Methods and
Programs in Biomedicine 71, 155 - 163
Heinze, G., Schemper, M. (2002): "A solution to the problem of separation in logistic regression", Statistics in Medicine 21, 2409 - 2419
Heinze, G., Schemper, M. (2001): "A solution to the problem of monotone likelihood in Cox regression", Biometrics 57, 114 - 199
Copas, J. B. (1997): "Using regression models for prediction: shrinkage and regression to the mean", Statistical Methods in Medical Research 6, 167 - 183
Mittlböck, M., Schemper, M. (1996): "Explained variation for logistic regression", Statistics in Medicine 15, 1987 - 1997
Schemper, M. (1993): "The relative importance of prognostic factors in studies of survival", Statistics in Medicine 12, 2377 - 2382
Maddala, G. S. (1983): "Limited-Dependent and Qualitative Variables in Econometrics", Cambridge University Press

These macros are provided at the github repository:

Please report bugs or make suggestions for enhancements directly at this github repository by creating corresponding issues.